MedleyDB: A Dataset of Multitrack Audio for Music Research

Synopsis

MedleyDB is a dataset of annotated, royalty-free multitrack recordings for noncommercial and academic research. MedleyDB was curated primarily to support research on melody extraction, addressing important shortcomings of existing collections. For each song we provide melody f0 annotations as well as instrument activations for evaluating automatic instrument recognition. The dataset is also useful for research on tasks that require access to the individual tracks of a song such as source separation and automatic mixing.

If you are an NYU student and interested in being a part of this project, please fill out this form.

For more information, please visit MedleyDB's main website:

http://marl.smusic.nyu.edu/medleydb